Although these results are in mice, not people, they raise an exciting possibility which can be tested in humans.

One of the surprises in this study, where we put some of the people through the same protocol five times, is that there is substantial variation with a person on how they respond day to day. That can be due to a lot of potential sources, including environmental factors, that we don't understand.

Up until now we have chronically underestimated the very substantial differences in the way individuals respond to the same dose of the same drug.

Lots of genes could be important -- genes that determine how you absorb a drug, variations in one gene involved in breaking down the drugs, variations in the way that you excrete a drug.

I think aspirin would do exactly what you might expect -- attenuate, but not abolish the risk.

We've all seen people get hired into the wrong jobs, but this system would almost eliminate that possibility and instead identify the best candidates available in any given applicant pool. You could call it the Rambo factor; you get one Rambo and you can wipe out several platoons quite easily.